Length of intestinal contact on nutrient-driven satiety
J. H. MEYER, Y. TABRIZI, N. DIMASO, M. HLINKA, AND H. E. RAYBOULD
Departments of Physiology and Medicine, University of California Los Angeles School of Medicine
and the West Los Angeles Veterans Affairs Medical Center, Los Angeles, California 90073
length of contact; load dependence; caloric equivalencies;
intestinal integration
FOOD INTAKES in rats are suppressed in a doseresponsive fashion by nutrients in intestines (21).
Responses are mediated by intestinal sensory mechanisms that are selectively responsive to digestive products (21). Sensors for this feedback are distributed in
small and large intestine and are about equally responsive in proximal and distal small gut (21). We observed
(21) that suppression of food intake varied with load
(amount/min) of nutrient entering gut whether the
nutrients [oleate-monolein, dodecanoate, phenylalaninetryptophan, or maltose (data for the latter not shown)]
were infused at constant concentration, but varied
flows, or at constant flow, but varied concentrations.
Thus responses depended on nutrient loads (amount/
min) infused, independent of concentrations over the
ranges tested. We postulated that food intakes decreased with increasing loads of nutrients infused into
the gut (21) or instilled into the stomach (20) because
higher loads of nutrients contacted longer lengths of
small intestine to excite additive feedbacks from more
and more sensors along the bowel.
Load dependence of nutrient-driven feedback has
been shown to relate to length of contact for several
different nutrient-driven feedback responses for which
chemosensors are distributed along most of the intestinal length, just as they are for intestinal satiety. The
importance of length of contact has been verified repeatedly in dogs by showing that diverting incoming nutrients truncated the dose response for stimulation of
pancreatic secretion or for signalling inhibition of gas-
R1308
tric emptying. The fraction of maximal response in
these truncations was lower when nutrient was diverted from shorter segments, higher when nutrient
was allowed contact with longer segments, and maximal when nutrients had access to all intestine (5,
11–14, 23–25). Thus responses mounted with dose as
more and more sensors were contacted along increasing
lengths of bowel. Similarly, in rats (9), the secretion of
the lipoprotein apolipoprotein (Apo) A-IV [a putative
signal of hypothalamic satiety (20)] into intestinal
lymph increased with the load of fat entering the
duodenum, but this increase was the result of entraining each successive segment of small intestine, in turn,
to synthesize and maximally secrete Apo A-IV, as
increasing incoming loads saturated the more proximal
segments to reach more and more distal gut.
In 10- to 45-cm segments of canine or human gut, the
amount of monomeric nutrient absorbed within the
segment depended on load entering the segment and
was similar whether nutrient was infused at low concentration but at higher and higher flows or at constant
flow but increasing concentrations (8, 23, 25, 26). There
was a maximal rate of absorption along each centimeter within the segment, either because of saturation of
a transport carrier or because of a rate-limiting process
within the enterocyte (such as the export of fat into the
lymph). As incoming loads increased, loads soon exceeded absorptive capacity within the more proximal
centimeters of the segment. Eventually, the entire
segment reached a maximal absorptive capacity so that
further loading would result in spillage of more and
more nutrient beyond the segment. Thus length contacted (cm) by incoming nutrient varied with the load
entering the gut (1, 2, 11, 13, 23–25) and was an
integration of load (g/min) entering divided by absorption (g · min21 · cm21 ) of each succeeding centimeter of
bowel, whether nutrient was infused at constant flows
but increasing concentrations or at constant concentration but increasing flows.
Biological chemosensors exhibit concentration dependence in response to exciting molecules. When confined
to short lengths of proximal intestine to prevent recruitment of more distal sensors (13), oleate-monolein inhibited canine gastric emptying in a concentrationdependent fashion at least over the range of 3–27 mM.
If suppression of nutrient intake by duodenally perfused oleate was mediated by sensory mechanisms
confined to a short segment of duodenum and if signals
from such sensors were concentration dependent but
independent of length of gut contacted beyond the
duodenal sensory area, then suppression would increase with concentration of oleate infused, just as it
did when 20–80 mM oleate was infused at 12 ml/h into
rat duodenum (80 mM almost completely suppressed
intake). However, infusing 80 mM oleate under these
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
Meyer, J. H., Y. Tabrizi, N. DiMaso, M. Hlinka, and
H. E. Raybould. Length of intestinal contact on nutrientdriven satiety. Am. J. Physiol. 275 (Regulatory Integrative
Comp. Physiol. 44): R1308–R1319, 1998.—Chemosensors
throughout small bowel and colon inhibit food intakes when
contacted by monomeric nutrients. We postulated that caloriedependent inhibition of food intakes depended on additions of
feedbacks from sensors in proximal and distal bowel contacted after high intakes of nutrients. Therefore, we determined how feedback from sensors in proximal gut interacted
with feedback from simultaneously contacted sensors in
distal bowel and whether suppression of nutrient intakes by
intestinally perfused nutrients depended on length of gut
contacted. Suppression of food intakes by maltose simply
added to that from dodecanoate when both were present
together either in proximal or distal small bowel. When
dodecanoate was infused into proximal gut while maltose was
infused distally, suppression of intake was threefold higher
and was thus potentiated. Limiting contact of slowly absorbed lactose or oleate to 35 cm of jejunum nearly abolished
the satiating potencies each exhibited during access to whole
gut. The observations were consistent with our hypothesis.
SPATIAL INTEGRATION OF INTESTINAL SATIETY
disaccharidases (maltase, trehalase, and lactase) to
vary the length over which glucolytic products were
released to contact sensors along small intestine. We
assumed that only hydrolytic products in gut were
capable of signalling a suppression of food intakes (21).
Substrate-specific, brush-border glucosidases have differing distributions along rat small intestine (1, 2, 30).
Maltase abounds in proximal, middle, and distal thirds
(1, 30). In the proximal third of small bowel, trehalase
(a-1,1-glucosidase) is 25% as abundant as maltase; in
the middle third, it is 11% as abundant; and in the
distal third, it is nearly absent (1). Even if higher and
higher loads of duodenal maltose saturated jejunal
maltase, undigested maltose spilling even into ileum
would be readily hydrolyzed and could thus signal
satiety there. By contrast, trehalase is almost completely absent from the distal third of rat small intestine (1). In other words, because of the limited distribution of trehalase, the release of glucose from trehalose
(a-1,1-glucosylglucose) entering duodenum is pretty
much confined to the proximal two-thirds of small
bowel, regardless of load. Therefore, the potency of
trehalose should be significantly less than that of
maltose if the latter derived its full potency at high dose
from additive feedback along the entire small gut.
Lactase has a broader and much different distribution
than trehalase, as it is ,10% as abundant as maltase in
each third of rat small intestine (2, 30). In addition,
proximal colonic content is rich in lactase of bacterial
origin (2). Because lactase is even less abundant than
trehalase in proximal small intestinal mucosa, an even
larger fraction of moderate and even lower duodenal
loads of lactose (b-1,4-galactosylglucose) spill undigested into distal bowel, but there is enough lactase in
the middle and lower thirds of small bowel and, especially, in colon to ensure a length of contact by released
glucose and galactose over the entire small and large
intestine (2). Therefore, if the total length of intestinal
contact with glucose (and galactose) is an important
determinant of intestinal satiety, small to moderate
loads of duodenal lactose should prove to be more
satiating than similar loads of maltose, as there is
known to be a greater spread of glucolytic products
along the bowel during digestion of the former (2).
Thus our second approach compared dose responses
of duodenal trehalose with duodenal maltose and duodenal lactose with duodenal maltose.
Our third experimental test was to determine whether
dose-responsive satiety would be lessened if perfused
nutrients were confined to 35 cm of proximal small
bowel in a Thiry Vella loop instead of being given access
to whole small intestine. We first studied how confinement to 35 cm of bowel would affect satiation from
perfused lactose compared with maltose. If perfused
lactose indeed signalled from longer lengths of bowel
than more rapidly hydrolyzed and absorbed maltose,
then its ability to suppress nutrient intake should be
more profoundly reduced than that of maltose. We also
examined suppression of intakes by jejunally perfused
oleate-monolein vs. maltose. Because oleate-monolein
is more slowly absorbed than glucose or maltose [80
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
circumstances should have produced a similar maximal
response, whether the oleate was infused at 6 or 3 ml/h,
because the relevant sensors within the duodenum
would still be exposed to the slowly absorbed oleate and
would still respond maximally to the 80 mM concentration. Instead, virtually the same dose response was
observed when 80 mM oleate was infused at 3, 6, and 12
ml/h as when 20, 40, and 80 mM oleate was infused at
12 ml/h (21). Because luminal concentrations of fatty
acids, even after high-fat meals, are ,39 mM (27), the
80 mM concentration of oleate was probably supramaximal for fatty acid sensors. The various considerations
suggested that receptors along a considerable length
were maximally stimulated by the supraphysiologically
high concentration of 80 mM oleate and that the
response increased as successive feedback mechanisms
were entrained and maximally stimulated along increasing lengths of gut contacted by the increasing
loads of 80 mM oleate delivered with varied volume
flows.
It was relatively easy to test length dependency for
other types of feedback in dogs. Length dependency
was demonstrated by utilizing chronic fistulas for
reversible diversions at varying distances along canine
bowel. Such easy reversibility allowed randomized
comparisons between diversion at one point or another
versus full access. Easy reversibility also allowed us to
demonstrate, even in a few dogs, that satiety (evidenced by a decrease in food intake) significantly
increased when the length of gut contacted by lipolytic
products was extended from jejunum into ileum (3).
In contrast to canine gut, rat intestine is less muscular and much smaller in diameter, making it impossible
to surgically implant diverting cannulas without obstructing the small bowel. Consequently, we were forced
in this work to test length dependency in rats with less
direct methods. Three different techniques were utilized to test our hypothesis that suppression of food
intakes by intestinal nutrients varies with length of gut
(i.e., number of sensors) contacted.
Our first approach was to divide the delivery of a dose
between duodenum and midgut. We were never completely certain in our previous study (21) how much of
various loads of duodenally infused nutrients contacted
distal small intestine or even colon, but our observations suggested that, on duodenal infusion, the middle
and high doses of slowly absorbed oleate or the highest
dose of somewhat more rapidly absorbed dodecanoate
spread into the colon. If so, the high satiety from the
highest doses of oleate or dodecanoate might have
resulted from addition of feedbacks from jejunal, ileal,
and colonic sensors. In the first experiments, we wanted
to see whether inputs from proximal sensors would
somehow add together with feedback from sensors in
distal gut. The general idea was to determine whether
simultaneous infusions of half-doses into duodenum
and into midgut would evoke responses similar to or
even greater than suppressions of intake during infusion of a full dose alone entirely into duodenum.
In the second set of experiments, we took advantage
of the known regional distributions of three mucosal
R1309
R1310
SPATIAL INTEGRATION OF INTESTINAL SATIETY
mM oleate was absorbed from 65-cm segments of
canine proximal small intestine ,1/10th as fast as
1.0 M glucose (13)], we predicted that confining perfusate to a 35-cm loop would reduce whole gut suppression
of intake more for oleate than for maltose if length of
intestinal contact were important in these responses.
METHODS
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
We used the same experimental techniques in shamfeeding or naturally feeding rat models that we employed in
our earlier study. More complete information on how these
animals were prepared and how they were tested can be
found in the preceding paper (21). Both models had infusion
catheters placed in the duodenum (#2 cm from the pylorus)
and at mid small intestine (55 cm from the pylorus) through
which we could instill nutrients or control solutions in desired
amounts. The two models were previously shown to give
similar information, but the ability to test daily in the
sham-feeding model allowed more rapid completion of protocols, whereas the much greater experimental life of the freely
feeding model allowed repeated testings to be averaged across
day-to-day variations.
Animals were housed in individual cages in a room with a
natural light cycle. After recovery from surgery, sham-fed rats
were trained to eat their day’s ration from 1200–1700 and
were then deprived of food but not water from 1700–0900 the
next day. These animals were allowed to sham feed for 90 min
some time during 0900–1200. Naturally feeding rats were
trained to eat daily from 0900–1200 and were studied during
this feeding period. For the remaining 21 h, the naturally
feeding rats were deprived of food but not water. In either
model, constant intestinal perfusions with nutrient or control
solutions began at the start of the ingestive period and
continued for its full duration.
To study sham-feeding rats with Thiry Vella loops, the
following procedure was adopted. After conditioning to body
restraint, rats were prepared as standard sham-fed models
but only with one infusion catheter, 15 cm distal to the
pylorus. After recovery, the animals were trained for 14 days
to sham feed and then underwent 4–6 days of sham feeding
for determination on successive days of response to controls,
to oleate-monolein or lactose, and to maltose. Immediately
thereafter, the rats were fasted 20 h and then operated on
again. The jejunum was transected ,3 cm above the infusion
catheter, and the distal margin was closed with an inverting
suture of 6.0 Ethicon and buttressed with a patch of Marlex
(Bard Marlex mesh; Bard, Cranston, RI) sewn over the
surface of the closure. A second transection was made 35 cm
below the first to form a Thiry Vella loop of proximal jejunum.
The distal stoma of this loop was brought out through a stab
wound in the abdominal wall where it was sutured to the
skin. Intestinal continuity was reestablished by anastomosing the proximal margin at the first transection to the distal
margin of bowel from the second transection. When used,
perfusates entered the proximal end of the loop through the
polyethylene catheter and exited from the distal end out the
cutaneous stoma so that perfusates had access limited to 35
cm of gut. Rats were allowed to eat again one day later, and
sham feeding resumed on the fourth postoperative day. For 2
days, rats sham fed during perfusions of the Thiry Vella loops
with saline. Provided that each animal 1) continued to appear
healthy, 2) had resumed eating .70% of preoperative daily
intakes, and 3) had sham intakes .70% of preloop intakes,
the animals again underwent 2 or 3 consecutive days of
testing with oleate or lactose controls and oleate or lactose,
followed by another 2 days of maltose control and maltose.
Thus apparently healthy animals were retested over postoperative days 6–11. In these experiments, maltose was always
given after oleate or lactose as a positive control.
In some animals with Thiry Vella loops, we measured the
absorption of oleate from oleate-monolein when infused at
loads of 0.24, 0.48, and 0.96 mmol/h of oleate, at either fixed
concentration (80 mM) or fixed volume flow (12 ml/h). Nonfasted animals were perfused in restraint cages for 120 min;
they were perfused with oleate-monolein during the first 90
min and then with a wash of 0.15 M NaCl at 12 ml/h for the
last 30 min. Effluents from the loops were collected in catch
basins beneath the restraining cages, acidified with HCl, and
extracted two times with equal volumes of 5:4 petroleum
ether-ethanol. Extracts were evaporated to dryness, and
redissolved in absolute ethanol, so that their content of
recovered fatty acid could be determined by titration with 0.1
N NaOH to pH 9.50.
Most perfusates were made isosmolar by adding NaCl as
needed to achieve 300 mosmol/kgH2O. For solutions of carbohydrates, osmolalities were restricted to a maximum of 400
mosmol/kgH2O for duodenal instillates and 800 mosmol/
kgH2O for midgut infusions, because we found in other
experiments (21) that solutions of NaCl alone markedly
inhibited intakes at osmolalities above, but not below, 500
mosmol/kgH2O in duodenum and above, but not below, 800
mosmol/kgH2O in ileum. Oleate plus monolein were emulsified with 10 mM taurocholate at pH 7 in a 2:1 molar ratio
(oleate:monolein). Control solutions were NaCl plus 10 mM
taurocholate. The source of crude taurocholate was desiccated
ox bile (Sigma Chemical, St. Louis, MO), which also contains
lecithin. Most solutions at 21°C were instilled at pH 7, but
solutions of dodecanoate (C-12) were at pH 8.1. In tests of
maltose alone vs. maltose plus C-12, control solutions for
maltose alone were NaCl plus NaHCO3 (pH 8.1); for tests of
maltose plus C-12, control solutions (no maltose) containing
C-12 were NaCl plus C-12.
Except in animals having Thiry Vella loops, doses or
treatments were scheduled in Latin square designs. In most
tests, two-way ANOVAs were used to determine whether
there was a significant treatment effect. If so, we used linear
contrasts to compare individual treatments. If full doseresponse curves were generated, dose-responsive effects were
examined by linear regression of values in each animal, with
0 dose 5 control and successive doses (given in geometric
progression) arrayed arithmetically as doses 1, 2, and 3.
Significance (difference from 0 by t-test) of slopes (g ingested/
unit dose) was then computed from the mean and SE of
individual slope values from all animals. Sometimes the
slopes obviously differed between one set of perfusates and
another. At other times, slopes were similar, but one dose
response was obviously much above the other. To avoid
possibly spurious statistical significance from multiple, pointby-point comparisons (4), we first made global comparisons of
slope and additionally of sums of responses from all doses of
each perfusate. These comparisons were made by paired
t-tests (for 2 treatments) or by ANOVAs (for 3 or more
treatments). If significant differences in global comparisons
were detected by these means, then individual doses along
the dose response to one perfusate were compared with the
corresponding doses of the other perfusate to determine the
most likely origin of the global differences.
Because of the possibility that Thiry Vella loops would lose
responsiveness over time, experiments were scheduled in
these animals so that high-dose maltose was always given
after oleate-monolein or lactose as a kind of positive control.
These experiments were designed in couplets so that oleatemonolein (or lactose) was first alternated among animals
SPATIAL INTEGRATION OF INTESTINAL SATIETY
Table 1. Redistribution of the load
of dodecanoate on food intakes
Site
Infusion
Duodenum
Midgut
Duodenm 1
midgut
Duodenum
Midgut
80 mM @ 6 ml/h
80 mM @ 6 ml/h
80 mM @ 6 ml/h 1
80 mM @ 6 ml/h
80 mM @ 12 ml/h
80 mM @ 12 ml/h
Load
Amount
Eaten, g
0.48 15.1 6 1.2
0.48 15.2 6 0.8
0.96 7.5 6 2.0*
0.96
0.96
5.9 6 1.2
7.1 6 0.4
Difference From
Control, g
1.0 6 1.5
0.9 6 1.1
8.6 6 2.0
10.3 6 1.3
9.0 6 0.7
with its corresponding control. After the first 2 days, when
every animal had received nutrient and control perfusions,
each animal was perfused for two more days with maltose or
maltose controls in an order that alternated among animals.
RESULTS
Interactions within or between segments. The first
experiment was designed to determine whether suppression of food intakes from 0.48 mmol/h of dodecanoate
(C-12) infused into duodenum of naturally feeding rats
would add to suppression from dodecanoate infused
simultaneously into midintestine. Six naturally feeding rats were studied, and the order of perfusions
(Table 1) was varied among the rats. Simultaneous
infusion of 0.48 mmol/h of C-12 into duodenum and into
midgut (each as 80 mM C-12 at 6 ml/h) increased the
suppression of food intake by more than eightfold over
the suppression observed with the 0.48 mmol/h dose
infused singly at one site or the other (P , 0.001, Table
1). The response to 0.48 mmol/h of C-12 simultaneously
at the two sites was about the same as to 0.96 mmol/h
(as 80 mM at 12 ml/h) singly into duodenum or into
midintestine (Table 1).
We wished to determine whether two different nutrients (maltose and C-12) would add their suppressive
effects when perfused together into the same segment
of bowel, whether jejunum or ileum. Two groups of
sham-feeding rats were used: one for duodenal infusions and the other for infusions into midintestine. In
each experiment, one-half of the animals received
maltose alone (50–400 mM at 12 ml/h) first and then
maltose plus C-12, whereas the other one-half was
perfused with maltose plus C-12 first and then with
maltose alone. Doses of maltose were randomized in a
Latin square design.
When maltose (50–400 mM) or maltose (50–400 mM)
plus C-12 (30 mM) was infused into duodenum, feeding
decreased as the dose of maltose increased (Fig. 1A).
Sham-feeding responses to maltose plus C-12 were
numerically, but not statistically, greater than to maltose alone at the lower doses, but at the highest dose of
maltose, the responses to the two solutions were the
same. Neither the slopes nor the sums of the responses
were significantly different (paired t-tests).
The second group of nine rats was infused at midgut,
with the concentration of C-12 increased to 40 mM to
offset a slightly weaker effect of low doses of C-12 at
midgut compared with duodenum (21). Each type of
nutrient solution (those with maltose alone or those
with maltose plus C-12) was infused at 12 ml/h and was
Fig. 1. Effect of combinations of dodecanoate (C-12) with maltose vs. maltose alone on sham feeding intakes (means
6 SE) in three different groups of rats. A: duodenal perfusions of maltose vs. maltose 1 30 mM C-12 at 12 ml/h in 10
animals. B: midgut infusions of maltose or maltose 1 40 mM C-12 at 12 ml/h in 9 animals. C: duodenal infusion of
NaCl-NaHCO3 or 60 mM C-12 at 6 ml/h while saline or 0.1–0.8 M maltose was infused into midgut of 9 rats at 6 ml/h
into each site.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
Data are means 6 SE from 6 animals. Effect of redistributing a full
load (0.96 mmol/h) of dodecanoate (C-12) at duodenum or midgut
alone to one-half load each at duodenum and midgut simultaneously.
* Simultaneous infusions of 0.48 mmol/h into duodenum and midgut
resulted in significantly (P , 0.001, ANOVA) less food intake than
during 0.48 mmol/h singly into duodenum or into midgut, but intakes
during dual 0.48 mmol/h simultaneously into the two sites were not
different from 0.96 mmol/h alone into duodenum or midgut.
R1311
R1312
SPATIAL INTEGRATION OF INTESTINAL SATIETY
Table 2. Maltose vs. maltose 1 C-12
S Responses
Slopes
Perfusion Mode
n
Maltose
Maltose 1 C-12
Maltose alone
Maltose 1 C-12
Duodenum (12 ml/h)
Midgut (12 ml/h)
Duodenum (6 ml/h) 1 midgut (6 ml/h)
10
9
9
210.45 6 3.08
28.79 6 2.43
27.04 6 2.63
210.23 6 2.00
212.00 6 1.93
26.85 6 1.89
336 6 35
300 6 21
276 6 40
293 6 40
257 6 24*
133 6 19†
Data are means 6 SE and depict slopes of linear regressions from individual animals for dose responses and sum of responses (SResponses)
to control plus four doses of maltose (or maltose 1 C-12) from individual animals in three different groups of 9–10 rats (number of rats
indicated by n). Slopes of the dose responses within each group were not significantly different (paired t-tests) whether the maltose was
perfused alone or with dodecanoate (C-12), but the sums of responses were significantly less when maltose was perfused with C-12 when 0.36
mmol/h of C-12 were infused proximally while maltose (0.6–4.8 mmol/h) was infused distally. * P # 0.05 and † P , 0.005, paired t-test vs.
SResponses to maltose alone in the same group of rats. Although SResponses to maltose alone did not significantly differ among the three
groups of rats in the three experiments (one-way ANOVA), the SResponses to maltose 1 C-12 were significantly different (P , 0.005) across the
three experiments.
bility, we performed an analogous experiment in one set
of 11 naturally feeding animals. Over four experimental days, these animals were perfused with a low dose of
maltose alone into midgut, a low dose of dodecanoate
alone into duodenum, the dodecanoate plus the maltose
together into duodenum, or the dodecanoate into duodenum simultaneously with the maltose into midgut.
When dodecanoate or dodecanoate plus maltose was
infused into duodenum, the midgut was perfused with
400 mosmol/kgH2O NaCl, and when maltose alone was
infused into midgut, the duodenum was perfused with
0.1 M NaHCO3 plus 0.5 M NaCl (Table 3). In the four
tests, duodenum and midgut were thus perfused simultaneously each at 6 ml/h with nutrient or saline. The
low doses of maltose alone or C-12 alone only slightly
suppressed daily intakes of rat chow. The maltose plus
C-12 together into duodenum suppressed food intakes
somewhat (but insignificantly) more than maltose alone
into midgut or C-12 alone into duodenum. However,
there was a more marked and significant (P , 0.01)
suppression of intake when C-12 was perfused into
duodenum while maltose was instilled at midgut, compared with maltose plus C-12 together into duodenum
(Table 3), even though the caloric loads were the same
during these last two treatments.
Dose responses to three disaccharides. This series of
experiments began with a comparison of dose responses to duodenal trehalose vs. duodenal maltose.
Table 3. Potentiation between proximal and distal gut
Perfusion Mode
kcal
Infused/3 h
Amount
Eaten, g
Distal maltose
Proximal C-12
Proximal maltose 1 proximal C-12
Proximal C-12 1 distal maltose
10.4
1.9
12.3
12.3
14.5 6 1.0
14.6 6 0.9
13.9 6 1.3
9.2 6 1.2*
In a Latin square design, 11 naturally feeding rats were perfused at
6 ml/h during daily 3-h feeding periods on four test days with distal
maltose alone at 2.4 mmol/h, proximal dodecanoate (C-12) alone at
0.36 mmol/h, proximal maltose together with proximal C-12, or
proximal C-12 plus distal maltose. Bowel segments not perfused with
nutrient were perfused with saline control solutions, so that both
proximal and distal bowel were perfused at 6 ml/h in all 4 tests.
Average daily intakes of rat chow (3.4 kcal/g) in these animals while
not perfused was 15.3 6 0.7 g. * Significantly less (P , 0.01, 2-way
ANOVA) than during proximal C-12 plus proximal maltose.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
found to inhibit sham feeding in a dose-related fashion,
but the inhibitions were higher at all doses of maltose
plus C-12 than with maltose alone (Fig. 1B). The sum of
the responses (Table 2) to maltose plus C-12 was higher
than to maltose alone (P # 0.05).
In the preceding experiments, we tested interactions
between maltose and C-12 when they were distributed
in the same segment of either proximal or distal small
bowel. In a third set of nine sham-feeding rats, we
wished to determine how maltose and C-12 would
interact when they were distributed separately, with
C-12 in proximal bowel and maltose in distal small
intestine. Therefore, we instilled 0.36 mmol/h of C-12
(that is, 60 mM C-12 at 6 ml/h) or NaCl-NaHCO3
(control solution at pH 8.1) at 6 ml/h into duodenum
while perfusing at midgut at 6 ml/h with 0–4.8 mmol/h
of maltose. Four animals first received maltose alone
into midgut plus saline-bicarbonate into duodenum
and then maltose into midgut plus C-12 into duodenum. The other five animals were perfused in the
reverse order.
There was a much greater interaction (Fig. 1C) than
in the previous two tests. Even the weak satiety
stimulus (6) of ileal 800 mosmol/kgH2O NaCl (control)
was markedly increased by the duodenal C-12 (vs. the
control duodenal perfusions with bicarbonate), and,
similarly, all responses to ileal maltose were markedly
enhanced by the duodenal C-12 when compared with
ileal maltose plus duodenal bicarbonate. Despite the
markedly increased interactions between stimuli during recruitment of sensors along the entire intestinal
length, the slope of the dose response to the ileal
maltose was not affected (Fig. 1 and Table 2). A one-way
ANOVA of the sums of responses (Table 2) indicated
that the three groups of rats responded similarly (P .
0.25) to maltose plus bicarbonate but differently (P ,
0.025) to maltose plus C-12, almost entirely because of
the much greater satiety with the duodenal C-12 plus
ileal maltose.
Because our sham-feeding animals had useful experimental lives of only ,3 wk, we were forced to use three
different sets of rats to complete the above experiments
(Fig. 1). Consequently, there was the possibility that
the third set of rats somehow differed from the others to
give a spurious impression of potentiation between
proximal and distal segments. To eliminate this possi-
SPATIAL INTEGRATION OF INTESTINAL SATIETY
duodenal maltose in 10 naturally feeding rats. One-half
of the animals received lactose (100–400 mM) first, and
the other one-half received maltose (100–400 mM)
first. Doses were given to individual animals under a
Latin square design. Both disaccharides (Fig. 2C)
inhibited intakes in a dose-responsive fashion (P ,
0.0005). The slopes of the two dose-response curves
(22.13 6 0.28 for maltose vs. 22.39 6 0.22 for lactose)
did not significantly differ because the responses to the
highest doses of each disaccharide were the same, but
the sum of responses was significantly (P , 0.025)
lower for lactose because each of the lower loads of
lactose, but especially the 2.4 mmol/h dose (P , 0.005),
inhibited food intakes more than the corresponding
loads of maltose.
The experiment was repeated to compare lactose
with maltose when each was infused into midgut in
another group of nine rats. The low and middle loads of
midintestinal lactose were no longer more effective
than isocaloric loads of maltose (Fig. 2D). The differences between the two sugars observed on duodenal
infusion disappeared on infusion into midgut, when
contact with the proximal one-half of small gut was
bypassed by the midintestinal infusion. The result was
consistent with the idea that lower loads of duodenal
lactose were more effective than isocaloric maltose
because more slowly hydrolyzed products from the
duodenal lactose contacted a longer total length of
Fig. 2. Effect of infusing various disaccharides into
duodenum (A–C) or mid small intestine (D) of
different groups of rats at 12 ml/h. Solutions of
glucose or galactose (and controls) were perfused at
24 ml/h, instead of 12 ml/h. Most of the animals that
were perfused with lactose (C) had also been tested
separately with glucose and with galactose (B). N,
no. of rats tested. k and ), Control intakes during no
perfusion.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
Five naturally feeding rats received duodenal maltose
(100–400 mM at 12 ml/h) first and then duodenal
trehalose (100–400 mM), and four rats received the
reverse. Doses of each sugar were randomized in a
Latin square design. There was significantly (P ,
0.005) less dose-responsive suppression of food intakes
by duodenal trehalose (Fig. 2A). The slope of this dose
response (20.99 6 0.21) was significantly (P , 0.0005,
paired t-test) less negative than that from maltose
(22.37 6 0.21) in the same animals, and, similarly, the
sum of responses to trehalose differed (P , 0.025) from
that to maltose. Both differences derived from a much
lower suppression after the 4.8 mmol/h dose of trehalose compared with maltose, as the low and moderate
loads of trehalose gave responses similar to isocaloric
loads of maltose.
Lactose is hydrolyzed in the intestine to glucose plus
galactose, whereas maltose yields two glucoses. Before
comparing lactose with maltose, we had to establish
that glucose and galactose were equipotent in inhibiting food intakes. The dose responses to each duodenally
infused sugar were compared in 12 naturally feeding
animals, with one-half of the animals receiving the
glucose (100–400 mM at 24 ml/h) first and the other
one-half the galactose (100–400 mM at 24 ml/h) first
(Fig. 2B). The two hydrolytic products of lactose proved
to be equally satiating. Next, we compared doseresponse curves from duodenal lactose with those of
R1313
R1314
SPATIAL INTEGRATION OF INTESTINAL SATIETY
Fig. 4. Effects of a jejunal load of oleate-monolein (0.96 mmol/h or 4.0
kcal/h) and a jejunal load of maltose (4.8 mmol/h or 6.9 kcal/h)
compared with individual control perfusions in 9 rats when access
was given to entire intestine (top) and again, later, when access was
confined to a 35-cm Thiry Vella loop of jejunum (bottom). Control
perfusions for oleate-monolein were 10 mM taurocholate in 300
mosmol/kgH2O saline and, for maltose, 400 mosmol/kgH2O saline. All
infusions were at 12 ml/h. ** P , 0.01 vs. control.
Fig. 3. Effects of 4.8 mmol/h (6.9 kcal/h) jejunal loads of lactose and of
maltose vs. jejunal saline in 11 rats. Each sugar was paired with and
compared with its own saline control (at 12 ml/h) when access was
given to entire intestine (top) and again, later, when access was
confined to a 35-cm Thiry Vella loop of jejunum (bottom). ** P , 0.01
and * P , 0.05 vs. control.
control (Fig. 4), and each nutrient inhibited significantly (P , 0.01, paired t-test). When the same perfusates were later confined to 35 cm of the previously
perfused jejunum in the Thiry Vella loops, the oleatemonolein no longer significantly inhibited intakes,
whereas the maltose still inhibited (P , 0.01) nearly as
well as it did when given access to the whole gut.
To confirm that oleate was in fact absorbed, we
measured how much perfused oleate we could recover
in the effluents from loops of 12 animals (Fig. 5) when
three loads of oleate-monolein were infused at varying
concentration, constant flow or constant concentration,
or varied flow. It was apparent that the oleate was
absorbed about as well when infused at varying concentrations vs. varying flows (Fig. 5). Thus the sums of
absorbencies (926 6 109 µmol/h absorbed during varying concentrations at 12 ml/h vs. 1,014 6 103 µmol/h
absorbed during varying flows at 80 mM concentrations) were not significantly different (paired t-test). If,
however, the common dose (960 µmol/h) was removed
from these sums, then the sum of absorbencies during
the lower two loads was significantly (P , 0.005) higher
during perfusions with 80 mM oleate at varied flows.
In these experiments, sham feeding was measured in
a total of 28 rats in which perfusions of maltose vs.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
proximal plus distal gut than did glucose released from
rapidly hydrolyzed maltose. This idea was tested further in Limiting access to 35 cm of jejunum.
Limiting access to 35 cm of jejunum. Eleven shamfeeding animals survived to be tested with lactose and
with maltose before and after creation of a jejunal Thiry
Vella loop (each was instilled at 4.8 mmol/h or 6.9 kcal/h
at 12 ml/h; Fig. 3). When given access to the whole gut
(before creation of the Thirty Vella loops), the isocaloric
loads of lactose and of maltose potently (P , 0.01)
inhibited sham feeding intakes compared with control
perfusions with 400 mosmol/kgH2O saline, and each
sugar inhibited about equally, just as it had at this high
dose in the naturally feeding rats described above.
When confined to 35 cm of jejunum, the inhibition from
each sugar (difference from control) dropped, significantly with lactose (P , 0.05 ANOVA) but not significantly with maltose.
Another nine animals survived in good health after
the creation of the Thiry Vella loops to be tested with
oleate-monolein and with maltose. When given access
to whole gut, the oleate-monolein (0.96 mmol/h or 4.0
kcal/h at 12 ml/h) reduced intakes (compared with its
300 mosmol/kgH2O saline plus taurocholate control) at
least as much as maltose did (4.8 mmol/h or 6.9 kcal/h)
when compared with its 400 mosmol/kgH2O saline
SPATIAL INTEGRATION OF INTESTINAL SATIETY
saline were carried out before and after creation of
35-cm loops (11 animals for experiments with lactose, 9
with oleate, plus 8 more for testing absorption of
oleate). Although there was quite a bit of variation
among animals, it was, nevertheless, clear that limiting the maltose to the loops also reduced the suppressive effect that the maltose had during its access to the
whole gut (Table 4).
DISCUSSION
Intestinally perfused nutrients inhibited food intake
(21, 32) whether infused into proximal duodenum
(within 2 cm from the pylorus), proximal jejunum (15
cm from the pylorus in the present studies), mid small
intestine [55 cm from the pylorus (6, 21)], distal ileum
[95 cm from the pylorus (32)], or cecum [125 cm from
the pylorus (21)]. The cecal infusion site was not tested
for dose responsiveness, but all of the other sites
demonstrated dose-responsive inhibitions of similar
magnitudes. It was unknown in these studies how far
perfused nutrients spread still unabsorbed distally
Table 4. Sham feeding intakes during access of
perfusates to whole gut vs. to 35-cm Thiry Vella loops
Saline controls
Maltose (4.8 mmol/h)
Whole Gut
Thiry Vella Loop
39.6 6 3.5
16.5 6 2.3
43.9 6 3.7
26.0 6 3.8*
Data are means 6 SE. Units are ml drunk in 90 min in 28 rats.
Intakes during control perfusions with 400 mosmol/kgH2O saline did
not differ statistically, whether the saline had access to the whole gut
or was confined to the loops. All perfusions were delivered at 12 ml/h.
* Different from maltose during access to the whole gut (P , 0.05,
2-way ANOVA).
from the infusion sites. Nevertheless, patterns of diarrhea during various perfusions suggested that higher
satiating doses of duodenal oleate-monolein or lactose
spilled into colon, still not completely absorbed by small
gut. Thus the 0.96 mmol/h duodenal dose of oleatemonolein frequently gave rise to diarrhea on 3-h perfusions (exact tally not kept); 75% of 3-h duodenal
perfusions with 4.8 mmol/h of lactose resulted in diarrhea vs. 0% of duodenal perfusions with isocaloric
maltose. These patterns are quite consistent with the
fact that lactose and oleate are known to be absorbed
much more slowly than maltose (2, 5, 13). Indeed, our
studies of oleate absorption (Fig. 5) indicated that some
of the 0.96 mmol/h duodenal dose must have reached
the colon.
From these perspectives, it was quite reasonable to
postulate that the high inhibition of nutrient intakes by
higher doses of duodenal oleate-monolein or lactose
resulted from the addition of inhibitory inputs from
sensors known to be present and thought to be contacted along small bowel and colon. This idea was
reenforced by our observations that inhibitions from
the highest doses of each of these two nutrients were
diminished on infusion into midgut compared with
infusions into duodenum, that is, response diminished
when proximal jejunum was excluded from the perfusion (for example, see Ref. 21 and Fig. 2, D vs. C).
Finally, there were several precedents for the idea that
increasing satiety from increasing duodenal loads of
satiating nutrient resulted from contact of still unabsorbed nutrient with increasing numbers of additive
sensors along the gut length; such length dependence
had been demonstrated previously for nutrient-driven
stimulation of pancreatic secretion (5, 23–25), inhibition of gastric emptying (11–13), and inhibition of small
bowel transit (14). We believe that the various observations herein provide convincing evidence for the verity
of this hypothesis.
Simultaneous infusions into duodenum and midgut.
Clearly, suppression of food intake from the 0.48 mmol/h
of C-12 into duodenum combined with suppression
from 0.48 mmol/h of C-12 into midgut to give a high
degree of suppression similar to that observed when
0.96 mmol/h of C-12 was infused into duodenum (Table
1). There are only two likely explanations. 1) The 0.96
mmol/h dose of duodenal C-12 was only absorbed
one-half in the proximal gut so that one-half of this
duodenal dose spilled into distal bowel to excite satiety
sensors along the full length of contact. These sensors
in intestinal mucosa mediated the satiety response and
their inputs added together. Thus the additive response
would not be different whether one-half of the duodenal
load of 0.96 mmol/h reached distal intestine still unabsorbed or whether 0.48 mmol/h were infused simultaneously into duodenum and into midgut to expose a
similar total number of receptors along the gut. 2)
Alternatively, the inhibition of food intake was mediated postabsorptively by sensors in liver, fat depots, or
brain. No matter where the 0.96 mmol/h of C-12 was
delivered, the entire load reached these postabsorptive
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
Fig. 5. Absorption of oleate by 35-cm loops of jejunum in 12 rats
during infusions of oleate-monolein at 0.24, 0.48, and 0.96 mmol/h
given at constant flow (12 ml/h), variable concentration or at constant
concentration (80 mM), variable flow. Oleate was infused for 90 min.
Per hour absorptive rates were calculated by subtracting oleate
recovered from oleate infused over the 90 min and dividing this
difference by 1.5.
R1315
R1316
SPATIAL INTEGRATION OF INTESTINAL SATIETY
release of glucose to the proximal small intestine); 2)
low and moderate loads of duodenal lactose were more
effective than isocaloric loads of duodenal maltose
(presumably because even low loads of duodenal lactose
were known to spread glucolytic products farther along
the gut than isocaloric loads of maltose); and 3) by
contrast, lactose infused into midgut was not more
effective than maltose (presumably because the greater
spread of lactose on duodenal infusion was eliminated
by removing one-half of the small intestinal length
from the perfusion).
Because of the underlying assumption that the sugars signalled only via released glucolytic products, the
fact that the three predictions proved to be correct was
only weakly supportive of the idea that intensity of
feedback varied with length of gut contacted by glucolytic products. However, the argument was considerably strengthened by the additional observation (Fig. 3)
that confining lactose to a 35-cm loop of jejunum
abolished its suppressive effects observed previously in
the same animals when the lactose had access to
virtually the entire gut. That the efficacy of lactose was
more reduced by this confinement than the corresponding efficacy of maltose in the same animals is more
consistent with the idea that both disaccharides satiated through the release of their constituent sugars but
not consistent with the alternative, that satiation was
mediated by sensors specific to dimeric lactose and
dimeric maltose. If the alternative was correct, then
sensors specific to duodenal lactose must have been
either more responsive and/or more numerous in proximal bowel than putative sensors to maltose in order to
explain the higher efficacy of duodenal, but not midintestinal, lactose over maltose (Fig. 2, C and D). If that
were so, then confining both perfusates to proximal
bowel in the Thiry Vella loops should have diminished
the response to maltose more than to lactose; but, in
fact, the reverse was the case. Therefore, the higher
efficacy of duodenal lactose over maltose was the result
of a longer spread of lactose products along the gut.
Because we previously (21) observed good correlations between percent inhibition of intakes in shamfeeding vs. naturally feeding rats during intestinal
perfusions of various nutrients, we felt that it would be
informative to compare percentage (of control) intakes
during jejunal perfusion with maltose in sham-feeding
rats with Thiry Vella loops with percentage (of control)
intakes during duodenal trehalose in naturally fed
animals. A total of 28 animals received the high dose of
maltose before and after creation of Thiry Vella loops.
When the 4.8 mmol/h of maltose were given access to
the entire gut, sham-feeding intakes were 44 6 5% of
those during control perfusions with saline, but when
the maltose was later confined to the 35-cm Thiry Vella
loops, intakes during the maltose were 71 6 12% of
control intakes. These percentages before vs. after
creation of the loops differed significantly (P , 0.05,
paired t-test) from each other, although each was
nevertheless significantly lower than 100%. In comparison, during 4.8 mmol/h of duodenal trehalose given
access to the entire gut, food intakes in naturally
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
sensors because the C-12 was readily absorbed from all
intestinal sites exposed to it.
There are several arguments against the second
explanation, but the most revealing is our observation
(Fig. 1 and Tables 2 and 3) that simultaneous perfusions of C-12 into duodenum and maltose into midgut
produced a significantly (P , 0.01) greater inhibition of
nutrient intake than did perfusion of C-12 together
with maltose into the same segment of duodenum (or
midgut). If signalling were predominantly postabsorptive, then it should make no difference where along the
gut these absorbable nutrients were infused, but it did.
The other arguments against postabsorptive signalling
are less direct but still compelling: intravenous nutrients are not as potent and/or as rapidly effective as the
same amount of nutrients instilled directly into gut
[discussed more extensively elsewhere (21, 20)]; some
nonmetabolized or poorly metabolized glucose analogs
are nevertheless highly potent (21) on instillation into
gut. That slowly absorbed lactose was as potent as
maltose is a new argument against explanation 2
(discussed further, below). The sum of observations
makes explanation 2 implausible, so we are left to
conclude under explanation 1 that sensors along the
gut directly signal satiety, that the magnitude of the
satiety response is some sort of integration of inputs
from sensors along the length of bowel contacted by the
satiating nutrient, and that these combined inputs can
be potentiating if sensors are contacted simultaneously
in both jejunum and ileum (Fig. 1 and Tables 1–3).
Comparison of satiating efficacies of three disaccharides. These experiments were conducted under the
assumption that only glucolytic products released from
each disaccharide triggered the suppression of food
intakes, an assumption based on our previous observation (21) that acarbose, a competitive inhibitor of
intestinal glucosidases, significantly reduced the satiating effects of maltose infused at duodenum (not shown)
or at midgut. In other experiments (20), the noncompetitive inhibitor of lipase, orlistat, completely abolished
lipolysis and satiation from premeals of triglyceride, so
we reasoned that the statistically significant but only
partial blockade of the satiating effects of maltose was
the outcome of an only partial, competitive inhibition of
maltase by the acarbose. An alternative interpretation
might be that suppression of intakes by maltose was
signalled both by sensors triggered by maltose itself
and additional sensors triggered by glucose released
from the maltose.
Nevertheless, on the basis of this assumption [the
known distributions of the three disaccharidases in the
brush borders of enterocytes along the gut (1, 2, 30), the
known (1, 2) differences of spread of glucolytic products
from these three disaccharides along the gut (in turn,
the outcome of the content per cm of disaccharidase
activities), and the idea that suppression of nutrient
intakes varied with the length of gut contacted by
glucolytic products], we made three predictions (Fig. 2)
that proved to be correct: 1) duodenal trehalose was
less potent than duodenal maltose (presumably because the intestinal distribution of trehalase limited
SPATIAL INTEGRATION OF INTESTINAL SATIETY
On the other hand, lactose confined to 35 cm of
jejunum was less satiating than maltose similarly
confined (Fig. 3) in a situation in which each undoubtedly contacted the full length of the 35-cm loop. This
observation suggests that the 10-fold higher (2) rate of
hydrolysis and absorption of monomeric sugars from
maltose resulted in a slightly stronger (but not statistically significant and certainly much less than 10-fold)
stimulus from this fixed, 35-cm segment. Alternatively,
the mix of glucose and galactose from the lactose could
have been less potent than the two glucoses released
from the maltose, but this second idea seems unlikely
because duodenal glucose and galactose individually
were about equipotent (Fig. 2B). In other nutrientdriven, intestinal feedbacks, responses varied with
loads and/or concentrations of signalling nutrients
even when nutrients were confined to short, submaximal lengths of gut (5, 13, 23, 25). From all of these
considerations, we think it is likely that intestinal
satiety varies with an integration by mucosal sensors of
both total length of gut contacted and the intensity of
stimulus along each centimeter of contact.
Limiting access of oleate to 35 cm of jejunum. Our
studies on the absorption of oleate from perfusates of
oleate-monolein (Fig. 5) confirmed 1) that oleate was
slowly absorbed (5, 13) and 2) that its absorption was
similar (i.e., no significant difference between sums of
responses) whether varied loads were introduced at
high or low concentrations, respectively, at low or high
flows (8, 23, 25, 26). Although more oleate may have
been absorbed from the lower two loads when delivered
as 80 mM oleate at 3 or 6 ml/h compared with 20 or 40
mM oleate at 12 ml/h, it is probable that our inability to
collect small volumes of effluent adhering to body fur
and restraint cages resulted in a lower percentage
recovery of the smaller volume perfusates, giving, to
some degree, a distorted impression of higher absorption from the 80 mM oleate. On the other hand, it may
also be that the 80 mM concentration actually enhanced absorption by speeding diffusion into absorptive cells.
Sixty to seventy percent of the oleate was absorbed
from each of the two lower doses, but only 50% of the
960 µmol/h dose was absorbed by the jejunal loop that
was one-third of gut length. Because ileum absorbs fat
at one-half the rate of jejunum (33), about one-sixth of
the 0.96 mmol/h of duodenal oleate probably escaped
small intestinal absorption to spill over into colon, a
conclusion that well accounts for patterns of diarrhea
seen during perfusions with oleate-monolein. Our rather
crude data suggest that the loops reached maximal
absorptive capacities between the 480 and 960 µmol/h
loads because absorption dropped from 70 to 50% of
load between these two doses. It is known that the
entire rat small intestine can transport triglyceride
(resynthesized after absorption of fatty acids and
monoglycerides) at a maximum rate of 200 µmol/h (29)
or ,400 µmol/h of fatty acid for the proximal one-half of
small intestine. During the 960 µmol/h loads, absorption of ,480 µmol/h of oleate was of this order of
magnitude. Thus we would conclude that absorption of
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
feeding animals were 75 6 4% of control intakes,
whereas intakes during isocaloric, duodenal maltose
were 40 6 4% of control in the same animals. Release of
glucose from trehalose is functionally confined by the
distribution of mucosal trehalase (1) to the proximal
67% of gut, but trehalase activity declines rapidly along
the middle third. Thus there was a similar order of
satiation by maltose confined to the proximal one-third
of gut and by trehalose functionally confined to about
the proximal one-half of bowel.
In our earlier study (21), comparisons of duodenal to
midintestinal maltose were conflicting; in 9 shamfeeding animals there were no significant differences
between duodenal and midintestinal maltose, whereas
in 10 naturally feeding animals, the slope of the
duodenal dose response was slightly but significantly
(P , 0.025) more negative. In the present experiments,
there were seven additional naturally feeding animals
that received both duodenal and midintestinal maltose.
Among all 17 animals the slope of the dose response to
duodenal maltose was significantly more negative than
the slope for midintestinal maltose, mainly because
intakes during the 4.8 mmol/h of duodenal maltose
were much lower (P , 0.005) than those during the
same dose delivered at midgut (as in Fig. 2, B vs. D).
Therefore, either 1) proximal sensors were more potent
or more numerous than distal sensors or 2) the duodenal maltose gave a larger response because the whole
gut participated, whereas the midintestinal perfusions
bypassed the jejunum to diminish the total number of
additive sensors contacted. Because restricting maltose
to 35-cm loops of jejunum in 28 sham-feeding animals
significantly reduced inhibition of intakes by the maltose, finding 2 appears to be the correct alternative.
In comparison with isocaloric loads of duodenal maltose, duodenal lactose given access to the whole gut was
more effective at lower loads and as potent at the
highest load (Fig. 2). Yet the absorption of component
glucose and galactose from lactose is known to be very
slow in comparison with absorption of glucose from
maltose because of the rate-limiting, slow hydrolysis of
the lactose by low amounts of mucosal lactase. Thus
overall absorption of lactose from the entire gut may
have been only one-tenth as fast as overall absorption
of maltose (2). That such slowly absorbed duodenal
lactose was as effective or more effective than isocaloric
maltose on inhibiting food intakes indicates that lactose signalled at the level of the gut and not from
putative postabsorptive mechanisms, which might depend on quantitative metabolism of absorbed sugars.
This idea differs from, but is analogous to, previous
observations that partially metabolized sugars (Dxylose) or nonmetabolized sugars (a-methylglucose,
3-O-methylglucose) with affinities for the glucose transporter were as potent as fully metabolizable, isocaloric
glucose (21). Both observations (with dimeric lactose
vs. maltose and with monomeric glucose analogs vs.
glucose) reenforce the notion that short-term signaling
of satiety by these carbohydrates is an intestinal, not a
postabsorptive, phenomenon.
R1317
R1318
SPATIAL INTEGRATION OF INTESTINAL SATIETY
supersaturate pancreatic lipase. As duodenal loads
increase, lipolysis does not keep pace and is completed
farther and farther downstream (18, 20). Thus the ratio
of fat to lipase entering the duodenum, in turn related
to the load of fat ingested (14, 16, 19, 20), may profoundly affect the intestinal distribution of and satiation from lipolytic products (14, 20). By contrast, amylysis is a very rapid process that is carried out jointly by
luminal amylases and by saccharidases in enterocytes.
For many dietary saccharides, hydrolysis is so rapid
that the distribution of released glucose is determined
by transport maxima of cells along the gut (1, 2, 7).
That intestinal distributions of products stem from
absorptive, rather than hydrolytic, capacities probably
accounts for the fact that easily digestible starch,
Polycose, maltose, and glucose were about equally
satiating, calorie for calorie, (15, 21), in a system in
which feedback inhibition of food intake depends heavily
on length of sensory contact with released glucose.
Nevertheless, some polymeric carbohydrates (10, 31)
may satiate differently because their hydrolysis is rate
limiting (as with lactose or trehalose). Satiating potencies of polymeric nutrients may therefore depend on the
specifics of their digestion and on their digestibility (10,
28, 31), as well as on their caloric content.
We are grateful to Kent Lloyd for coaching us on how to make
Thiry Vella loops in rats.
This work was supported by research funds from the Department
of Veterans Affairs.
Address for reprint requests: J. H. Meyer, Rm. 105, Bldg. 115, West
Los Angeles VAMC, 11301 Wilshire Blvd., Los Angeles, CA 90073.
Received 12 May 1997; accepted in final form 23 June 1998.
REFERENCES
1. Dahlquist, A,. and D. L Thomson. The digestion and absorption of maltose and trehalose by the intact rat. Acta Physiol.
Scand. 59: 111–125, 1963.
2. Dahlquist, A., and D. L. Thomson. The digestion and absorption of lactose by the intact rat. Acta Physiol. Scand. 61: 20–33,
1964.
3. Doty, J. E., Y. G. Gu, and J. H. Meyer. The effect of bile
diversion on satiety and fat absorption from liquid and solid
dietary sources. J. Surg. Res. 45: 537–543, 1988.
4. Elashoff, J. D. Down with multiple t-tests. Gastroenterology 80:
615–620, 1981.
5. Fink, A. S., and J. H. Meyer. Intraduodenal emulsions of oleic
acid augment acid-induced canine pancreatic secretion. Am. J.
Physiol. 245 (Gastrointest. Liver Physiol. 8): G85–G91, 1983.
6. Glick, Z. Intestinal satiety with and without upper intestinal
factors. Am. J. Physiol. 236 (Regulatory Integrative Comp. Physiol.
5): R142–R146, 1979.
7. Gray, G. M. Carbohydrate absorption and malabsorption. In:
Physiology of the Digestive Tract, edited by L. R. Johnson, M. I.
Grossman, E. Jacobsen, J. Christensen, and S. Schultz. New
York: Raven, 1981, p. 1063–1072.
8. Holdsworth, C. D., and A. M. Dawson. The absorption of
monosaccharides in man. Clin. Sci. (Colch.) 27: 371–379, 1964.
9. Kalogeris, T. J., K. Fukagawa, and P. Tso. Synthesis and
lymphatic transport of intestinal apolipoprotein A-IV in response
to graded doses of triglyceride. J. Lipid Res. 35: 1141–1151, 1994.
10. Leathwood, P., and P. Pollet. Effects of slow release carbohydrates in the form of bean flakes on the evolution of satiety and
hunger in man. Appetite 10: 1–11, 1988.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
oleate by the jejunal Thiry Vella loops was completely
normal and, therefore, that the loss of inhibition of
sham feeding on confining oleate to these loops was not
an artifact of disordered absorption.
We observed nearly a complete loss of inhibition of
food intakes when perfused oleate-monolein was confined to 35-cm Thiry Vella loops in sham-feeding rats
compared with potent inhibition observed in the same
animals when the perfused oleate had access to the full
intestinal length (Fig. 4). However, in the same animals, maltose was as effective when confined to the loop
as it had been when given access to the entire gut. Thus
suppression of food intakes by intestinally perfused,
slowly absorbed oleate appeared to be much more
dependent on full intestinal access than was maltose.
This result with oleate-monolein was consistent with
three previous observations. 1) We observed in dogs (3)
that promoting (via reversible exclusion of bile) the
spread of digested, but unabsorbed, fat from an initial
meal into ileum significantly suppressed intake of a
second meal taken 4 h later. 2) Duodenally perfused
oleate-monolein (21) produced an inhibition of food
intake in naturally feeding rats with a significantly
more negative slope to the dose response than did the
same solutions infused at midgut, that is, the inhibition
was significantly less when the most proximal 50% of
the small intestinal length was removed from contact
with the oleate during midintestinal perfusions. The
principal reason for the greater negative slope of the
duodenally perfused oleate-monolein was that inhibition from the highest dose (0.96 mmol/h) was significantly greater on duodenal perfusion. 3) Similarly,
resecting the distal 50% of small bowel significantly
reduced inhibition of intakes of naturally feeding rats
by duodenally perfused oleate. Loss of inhibition from
oleate-monolein after removal of 50% of distal bowel
was greater than loss of inhibition after bypassing the
proximal 50% of bowel during perfusions into midintestine, but the smaller difference during midintestinal
perfusions, with bypass of proximal bowel, probably
resulted from the fact that some ingested rat chow
emptied into the bypassed jejunum in unresected animals where it also inhibited intake. The present result
(that confining oleate-monolein to 35 cm of jejunum
nearly abolished its satiating properties) is entirely
consistent with all three previous observations; all
indicate that high satiation from fat depends on entraining additive inputs from sensors of fatty acids along the
full intestinal length. Complete loss of efficacy on
confinement to 35-cm loops simply reflected a more
severe shortening of total length of contact with small
and large bowel than encountered when jejunum was
bypassed or ileum was resected.
Monomeric and dimeric nutrients were used to demonstrate how intestinal distributions of nutrients may
affect satiety. With polymeric foods, intestinal distributions of satiating hydrolytic products are additionally
determined by digestion. Thus daily intakes of chow
could be significantly reduced in dogs just by reversibly
shifting entry of pancreatic enzymes to ileum (17).
Normally, triglycerides enter duodenum at rates that
SPATIAL INTEGRATION OF INTESTINAL SATIETY
22. Meyer, J. H., and R. S. Jones. Canine pancreatic responses to
intestinally perfused fat and products of fat digestion. Am. J.
Physiol. 226: 1178–1187, 1974.
23. Meyer, J. H., G. A. Kelly, L. J. Spingola, and R. S. Jones.
Canine gut receptor system mediating pancreatic responses to
luminal L-amino acids. Am. J. Physiol. 231: 669–677, 1976.
24. Meyer, J. H., L. W. Way, and M. I. Grossman. Pancreatic
bicarbonate response to various acids in duodenum of dog. Am. J.
Physiol. 219: 964–970, 1970.
25. Meyer, J. H., L. W. Way, and M. I. Grossman. Pancreatic
response to acidification of various lengths of intestine in dog.
Am. J. Physiol. 219: 971–977, 1970.
26. Modigliani, R., and J. J. Bernier. The absorption of glucose,
sodium and water by the human jejunum. Studies by intestinal
perfusion with a proximal occluding balloon and at variable flow
rate. Gut 12: 184–193, 1971.
27. Porter, H. P., and D. R. Saunders. Isolation of the aqueous
phase of human intestinal content during the digestion of a fatty
meal. Gastroenterology 60: 997–1007, 1971.
28. Tournier, A., and J. L. Sylvestre. Effect of physical state of a
food on subsequent intake in human subjects. Appetite 16:
17–24, 1991.
29. Tso, P., K. L. Buch, J. A. Balint, and J. B. Rodgers. Maximal
lymphatic triglyceride transport rate from rat small intestine.
Am. J. Physiol. 242 (Gastrointest. Liver Physiol. 5): G408–G415,
1982.
30. Tsubol, K. K., L. K. Kwong, K. Yamada, P. Sunshine, and O.
Kolodovsky. Nature of elevated rat intestinal carbohydrase
activities after high-carbohydrate diet feeding. Am. J. Physiol.
249 (Gastrointest. Liver Physiol. 12): G510–G518, 1985.
31. Van Amelsvoort, J. M. M., and J. A. Westrate. Amyloseamylopectin ratio in a meal affects postprandial variables in
male volunteers. Am. J. Clin. Nutr. 55: 712–718, 1992.
32. Woltman, T., and R. Reidelberger. Effects of duodenal and
distal ileal infusions of glucose and oleic acid on meal patterns in
rats. Am. J. Physiol. 269 (Regulatory Integrative Comp. Physiol.
38): R7–R14, 1995.
33. Wu, A. L., S. Bennett-Clark, and P. R. Holt. Transmucosal
triglyceride transport rates in proximal and distal rat intestine
in vivo. J. Lipid Res. 16: 251–257, 1975.
Downloaded from http://ajpregu.physiology.org/ by 10.220.32.246 on June 16, 2017
11. Lin, H. C., J. E. Doty, T. J. Reedy, and J. H. Meyer. Inhibition
of gastric emptying by glucose depends on the length of the
intestine exposed to the nutrient. Am. J. Physiol. 256 (Gastrointest. Liver Physiol. 19): G404–G411, 1989.
12. Lin, H. C., J. E. Doty, T. J. Reedy, and J. H. Meyer. Inhibition
of gastric emptying by acids depends on the titrateable acidity
and the length of gut exposed to acid. Am. J. Physiol. 259
(Gastrointest. Liver Physiol. 22): G1025–G1030, 1990.
13. Lin, H. C., J. E. Doty, T. J. Reedy, and J. H. Meyer. Inhibition
of gastric emptying by sodium oleate depends on the length of gut
exposed to the nutrient. Am. J. Physiol. 259 (Gastrointest. Liver
Physiol. 22): G1031–G1036, 1990.
14. Lin, H. C., X.-T. Zhao, and L. Wang. Fat absorption is not
complete by midgut but is dependent on load of fat. Am. J.
Physiol. 271 (Gastrointest. Liver Physiol. 34): G62–G67, 1996.
15. McHugh, P. R., and T. H. Moran. Accuracy of the regulation of
caloric ingestion in the rhesus monkey. Am. J. Physiol. 235
(Regulatory Integrative Comp. Physiol. 4): R29–R34, 1978.
16. Meyer, J. H., J. D. Elashoff, M. Domeck, A. Levy, D. Jehn, M.
Hlinka, R. Lake, L. S. Graham, and Y. G. Gu. Control of
canine gastric emptying of fat by lipolytic products. Am. J.
Physiol. 266 (Gastrointest. Liver Physiol. 29): G1017–G1035,
1994.
17. Meyer, J. H., J. D. Elashoff, J. E. Doty, and Y. G. Gu.
Disproportionate ileal digestion on canine food consumption: a
model for satiety in pancreatic insufficiency. Dig. Dis. Sci. 39:
1014–1024, 1994.
18. Meyer, J. H., Y. G. Gu, D. Jehn, and J. E. Doty. Factors which
affect the performance of lipase on fat digestion in a canine model
of pancreatic insufficiency. Pancreas 9: 613–623, 1994.
19. Meyer, J. H., M. Hlinka, D. Kao, R. Lake, E. Maclaughlin,
L. S. Graham, and J. D. Elashoff. Gastric emptying of oil from
solid and liquid meals: effect of human pancreatic insufficiency.
Dig. Dis. Sci. 41: 1691–1699, 1996.
20. Meyer, J. H., M. Hlinka, A. Khatibi, H. E. Raybould, and P.
Tso. Role of small intestine in caloric compensations to oil
premeals in rats. Am. J. Physiol. 275 (Regulatory Integrative
Comp. Physiol. 44): R1320–R1333, 1998.
21. Meyer, J. H., M. Hlinka, Y. Tabrizi, N. Dimaso, and H. E.
Raybould. Chemical specificities and intestinal distributions of
nutrient-driven satiety. Am. J. Physiol. 275 (Regulatory Integrative Comp. Physiol. 44): R1293–R1307, 1998.
R1319